impact of thermal stress on evolutionary trajectories of ... · daily ad libitum with live artemia...

Schade et al. BMC Evolutionary Biology 2014, 14:164http://www.biomedcentral.com/1471-2148/14/164

RESEARCH Open Access

Impact of thermal stress on evolutionarytrajectories of pathogen resistance in three-spinedstickleback (Gasterosteus aculeatus)Franziska M Schade*, Lisa NS Shama and K Mathias Wegner

Abstract

Background: Pathogens are a major regulatory force for host populations, especially under stressful conditions.Elevated temperatures may enhance the development of pathogens, increase the number of transmission stages,and can negatively influence host susceptibility depending on host thermal tolerance. As a net result, this can leadto a higher prevalence of epidemics during summer months. These conditions also apply to marine ecosystems,where possible ecological impacts and the population-specific potential for evolutionary responses to changingenvironments and increasing disease prevalence are, however, less known. Therefore, we investigated the influenceof thermal stress on the evolutionary trajectories of disease resistance in three marine populations of three-spinedsticklebacks Gasterosteus aculeatus by combining the effects of elevated temperature and infection with a bacterialstrain of Vibrio sp. using a common garden experiment.

Results: We found that thermal stress had an impact on fish weight and especially on survival after infectionafter only short periods of thermal acclimation. Environmental stress reduced genetic differentiation (QST) betweenpopulations by releasing cryptic within-population variation. While life history traits displayed positive geneticcorrelations across environments with relatively weak genotype by environment interactions (GxE), environmentalstress led to negative genetic correlations across environments in pathogen resistance. This reversal of geneticeffects governing resistance is probably attributable to changing environment-dependent virulence mechanisms ofthe pathogen interacting differently with host genotypes, i.e. GPathogenxGHostxE or (GPathogenxE)x(GHostxE) interactions,rather than to pure host genetic effects, i.e. GHostxE interactions.

Conclusion: To cope with climatic changes and the associated increase in pathogen virulence, host species requirewide thermal tolerances and pathogen-resistant genotypes. The higher resistance we found for some families atelevated temperatures showed that there is evolutionary potential for resistance to Vibrio sp. in both thermalenvironments. The negative genetic correlation of pathogen resistance between thermal environments, on theother hand, indicates that adaptation to current conditions can be a weak predictor for performance in changingenvironments. The observed feedback on selective gradients exerted on life history traits may exacerbate this effect,as it can also modify the response to selection for other vital components of fitness.

Keywords: Climate change, Infectious diseases, Vibrio tubiashii, Genotype x environment Interaction, Selection

* Correspondence: [email protected] Wegener Institute Helmholtz Centre for Polar and Marine Research,Hafenstrasse 43, 25992 List/Sylt, Germany

© 2014 Schade et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly credited. The Creative Commons Public DomainDedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article,unless otherwise stated.

mailto:[email protected]

Schade et al. BMC Evolutionary Biology 2014, 14:164 of 12http://www.biomedcentral.com/1471-2148/14/164

BackgroundPathogens are strong selective agents that can influencebiodiversity by regulating host population dynamics [1,2].Environmental conditions can amplify these effects [3],and many infectious diseases are temperature-dependent[4]. An increase in ambient temperature will affect thepathogen by, for example, enhancing its metabolismresulting in faster development and a higher number oftransmission stages per generation. Especially for manybacterial disease agents, population growth rates in-crease exponentially with increasing temperature [5,6].Since population growth rate closely resembles fitnessfor prokaryotes, improved pathogen fitness can also leadto a higher prevalence of diseases [7].Rising temperatures are also likely to stress the host or-

ganism, thereby increasing host susceptibility to diseases[8,9]. This also applies to marine organisms [10], whichare often highly vulnerable to climatic changes [11]. Ther-mal stress causes an alteration of the immune response,which generally results in a reduction of pathogen resist-ance [12]. Despite these multifaceted effects, the specificinteractions of increased thermal stress due to climaticwarming with infectious diseases are sparsely explored,particularly in marine ecosystems.The ecological and evolutionary consequences of ther-

mal stress-disease interactions will strongly depend onthe thermal tolerance and rate of adaptation of popula-tions to climate warming [13]. A high potential for accli-mation within and between generations and the capacityfor adaptation by selection of pathogen-resistant geno-types will ensure population viability and biodiversity[14,15]. Species with rapid turnover of generations arelikely to adapt fast enough to cope with the changing en-vironment [16]. For vertebrates, the potential for adapta-tion by microevolution is considered to be alarminglylow [17]. Standing genetic variation supplies the raw ma-terial for adaptation, and different genotypes will showdifferent reaction norms leading to genotype by environ-ment (GxE) interactions. GxE interactions as well as pheno-typic plasticity can potentially compensate for declines inmean fitness of populations with low evolutionary potential.Furthermore, it also has been shown that changing en-vironmental conditions can release otherwise crypticgenetic variation [18], which can consequently alter theevolutionary potential of a population.Here, we tested whether environmental change can

also alter the genetic components of resistance to patho-gen challenges, and whether such changes can feed backon selection gradients of other components of fitness.To do so, we set up a common garden breeding experi-ment using three populations of marine three-spinedsticklebacks Gasterosteus aculeatus and challenged theoffspring with an infection by a bacterial strain of Vibriosp. in two different temperature environments (ambient

and stressful). In general, bacteria of the genus Vibrio in-clude many facultative symbionts and pathogenic strains[19]. They are strongly temperature-dependent with ele-vated population sizes during summer [20,21]. Moreover,they are able to adapt rapidly to environmental changesdue to high genome plasticity including frequent muta-tion, recombination and lateral gene transfers [22], mak-ing them ideal candidates for our purpose to study host-parasite interactions in changing environments. Vibriotubiashii, in particular, has been described as a pathogenof bivalves [23,24], but is also known to harm severalmarine fish species [25,26], where Vibriosis can lead tosuperficial skin lesions, muscle necrosis and haemor-rhages which are likely to cause mortality [27].The three-spined stickleback represents an ideal model

system to study the evolutionary ecology of host-parasiteinteractions [28-30] especially in connection with envir-onmental change. These fish occur in shallow freshwaterand marine habitats with a temperature range from 4 to20°C [31], but show preferences for intermediate tempera-tures (15-18°C [32,33]). Our experimental temperature en-vironments were set to simulate the average summertemperature of coastal North Sea areas (i.e. 17°C) andelevated temperatures with reference to recent climatechange predictions for the North Sea (21°C) [34],which was shown to result in diminished growth ratesafter long-term exposure in one of the populationsused here [35].With this study, we estimated the effects of thermal

stress on pathogen resistance (i.e. survival after infection)and its interactions with other life history traits that canaffect an individual’s survival and reproductive potential.In a narrower sense, life history traits include gestationtime, age to sexual maturity, reproductive life span andnumber of progeny. In a wider sense, morphological traitssuch as length and weight can also be seen as life historytraits representing key maturational and reproductivecharacteristics, and we thus chose to also measure sizeand weight.Short-term fluctuations and extreme events like heat

waves that are predicted to increase in the nearer future[36] could have very different effects on populationswith different adaptive thermal optima. By additionallydetermining the within- and between-population gen-etic components of resistance and life history traits, wecan also project if environmental stress will modify theresults of population-specific evolutionary trajectoriesby releasing or absorbing genetic variance for traitsunder selection. The connection between environmen-tally modified pathogen-induced selection to its conse-quences for other components of fitness will help todeepen our understanding of how populations can orcannot cope with increasingly stressful and pathogenrich environments.


MethodsSampling and fish breedingDuring April and June 2011, adult marine sticklebackswere caught by dip netting in an oyster farm in Yerseke(Oosterschelde, The Netherlands, 51.49 N, 4.06 E) andin tidal channels on the islands of Texel (The Netherlands,53.16 N, 4.88 E) and Sylt (Germany, 55.03 N, 8.46 E). Be-tween 50–70 individuals were sampled from each locationand transferred in groups of 5–7 fish to 20 l aquaria (38 ×25 × 20 cm) equipped with permanent seawater flow-through at a constant temperature of 17°C. Between Julyand August, laboratory-bred F1 families were producedfor each population (20 families per population) using amodified North Carolina II breeding design (five replicatesof crosses between two dams x two sires i.e. four half-sibling families). Since not all of the crosses resulted insuccessful hatchings, we produced additional families (n =10 and using different parents) to increase sample sizeand achieve better estimates of within-population vari-ation for the Sylt population. Eggs were kept in aerated 1 lglass beakers until hatching. Hatchlings were then trans-ferred to 10 l flow through tanks (19 × 25 × 20,5 cm) witha water temperature of 17°C and a 16 h: 8 h light: darkcycle. In total, 14 families from Oosterschelde, 11 familiesfrom Texel and 24 families from Sylt hatched.One month after hatching, the density of each family

was reduced to 20 fish per 10 l aquaria. Fish were feddaily ad libitum with live Artemia sp. until 12 weeks ofage and afterwards with frozen chironomid larvae untilthe end of the experiment. During the early rearingphase, we lost several families due to electrical failure ofaeration pumps. These affected only newly hatched fam-ilies from all populations and did not lead to a systematicbias in mortality (χ22 = 0.584, P = 0.747). In the end, ninefamilies remained from the Oosterschelde population(nfish = 164), nine families from Texel (nfish = 192) and 13families from Sylt (nfish = 263) that had sufficient familysizes (20 fish per family) for the infection experiment.

Infection experimentEight months after hatching, families were separated intosmaller groups of 10 fish per 2,6 l aquarium (12 × 18,5 ×11,8 cm). In line with recent climate change predictionsfor the North Sea [34], half of each family (n = 10) wasexperimentally exposed to elevated temperatures by in-creasing the water temperature up to 21°C. The remaininghalf (n = 10) was kept at their rearing temperature of 17°Cas a control. After a four week thermal acclimation period,fish from both temperature treatments were injected ei-ther with 10 μl of a solution containing a bacterial strainof Vibrio sp. closely related to a pathogenic strain of Vibriotubiashii (107 cells per ml, 17°C n = 5, 21°C n = 5) or withthe same amount of phosphate buffered saline (PBS)medium as control (17°C n = 5, 21°C n = 5). Infected and

control fish were then held in ambient or elevated watertemperature without flow through and susceptibilitywas determined by monitoring induced mortality every12 hours for 10 days. Dead fish as well as fish remainingat the end of the experiment were weighed (wet mass tothe nearest 0.01 g) and measured (standard length tothe nearest 0.1 cm). All experiments were conducted inaccordance with German animal welfare laws (Permis-sion No. V312-72241.123-16).

Microsatellite genotypingTo measure drift-based neutral divergence between pop-ulations, each parental fish (n = 70) was genotyped at 15microsatellite loci. DNA was extracted from the caudalfin using the DNAeasy Blood and Tissue Kit (Qiagen,Hilden). The 20 μl multiplex PCR reactions consisted of4 μl of 5x PCR Flexi buffer, 2 μl 10 mM dNTP, 0.6 μl25 mM MgCl2, 0.1 μl 500 U GoTaq DNA Polymerase(all PCR chemicals from Promega, Mannheim), 5 pmolof each fluorescently labelled forward and unlabelled re-verse primer and 2 μl DNA template. Thermal cyclingfor PCR group 1 (containing markers 5196 HEX, 41706_FAM, 1125 6_FAM, 1097 NED and 7033 NED, [37])started with an initial denaturation step at 94°C for3 min followed by 30 cycles of 94°C for 1 min, 58°C for1 min and 72°C for 1 min. PCR group 2 (containingmarkers STN 18 HEX, STN 32 6_FAM, STN 75 HEX andSTN 84 NED) started with an initial denaturation step at94°C for 3 min followed by 30 cycles of 94°C for 45 sec,56°C for 45 sec and 72°C for 1 min. PCR group 3 (contain-ing markers STN 2 6_FAM, STN 170 HEX and STN 174HEX) and as well as PCR group 4 (STN 36 6_FAM, STN114 6_FAM and STN 167 HEX, [38]) started with an ini-tial denaturation step at 94°C for 2 min followed by 30 cy-cles of 94°C for 45 sec, 56°C for 45 sec, 72° for 45 sec anda final extension step at 72°C for 20 min.One μl of each PCR reaction was denatured in HiDi-

formamide (Life Technologies, Darmstadt) at 94°C for2 min. Fragments were separated using an ABI prism3100 XL (Life Technologies, Darmstadt) capillary sequen-cer. Genotype scoring was performed using GeneMarkersoftware (SoftGenetics, version 1.91), and neutral geneticdifferentiation (measured as FST) was calculated afterWeir and Cockerham [39] using the software GENETIX[40]. Significance of the observed estimates was calculatedby bootstrapping using 1000 random permutations of thedata. The range for the neutral marker-based differenti-ation in the QST-FST comparison was determined by usingthe extremes of the distribution of single-locus estimatesof FST for all 15 loci [41].

Statistical analysesWe used generalized linear mixed models (GLMM)using the MCMCglmm package [42] of the R statistical


environment [43] for all of our statistical analyses. First,we wanted to determine the effects of temperature andinfection as fixed effects on length, weight and survivalusing population and full-sibling family as random ef-fects. Density and age were entered as covariates in allmodels to account for differences among families priorto our experimental treatment. We fitted models as gen-eralized linear mixed models with length and weight asGaussian and survival as binomial response variables.Markov chains were run for 500’000 iterations and wekept every 100th value after removing 300’000 iterationsof burn-in to generate posterior distributions of randomand fixed parameters. Model fits were assessed by theirrespective Deviance Information Criterion (DIC) scores[44] including random effects. We used weak but in-formative priors of half the observed phenotypic vari-ance and examined posterior samples from the Markovchain for signs of autocorrelation. To test for GxE inter-actions, we additionally fitted random slopes for randomterms (singly and combined), and assessed model fit byDIC scores taking a better model fit for a full-siblingfamily random slopes model as an indication for GxEinteractions.Second, we used population and animal variance com-

ponents from environment and trait specific GLMMs tocalculate population and genetic differentiation in differ-ent environments. Divergence in quantitative traits (QST)was calculated as [45],

QST ¼ σ2Bσ2B þ 2σ2W

where σ2B and σ2W are the between- and within-population components of genetic variation. The animalvariance component was taken as the measure of within-population variance, while the population variance com-ponent was taken as a measure of between-populationvariation.We used a character state approach [46], treating life

history traits and resistance in the different thermal en-vironments as separate traits to calculate genetic correla-tions of traits across environments. Since individualscould only be assayed in one environment, we did notestimate covariance on the individual/unit level (byusing the idh(trait):units covariance structure imple-mented in MCMCglmm). Variance and covariance forcalculating genetic correlations (rG) were estimated onthe level of the pedigree by using the us (trait):animalvariance components. Genetic correlations across envi-ronments were calculated as the covariance betweentraits divided by the square root of the product of bothtrait variances. Significance of genetic correlations acrossenvironments was assessed by estimating the proportion

of estimates from the posterior distributions that over-lapped with zero.The effect of selection gradients imposed on life history

traits by bacterial infection was analysed using only datafrom infected fish in separate thermal environments.GLMMs were fitted with weight or length as response var-iables, survival status as a fixed effect, density as a covari-ate, and population and animal as random effects.

ResultsEffects of thermal stress on life history traits andpathogen resistanceThermal stress was visible in life history traits of stickle-backs after the acclimation and experimental period(Figure 1). Fish were on average smaller and lighter at21°C than at 17°C, but temperature only had significanteffects on weight (Table 1). Models fitting separate inter-cepts but common slopes for populations showed a su-perior fit, indicating that the effect of temperature onweight was not significantly different between popula-tions, whereas length and weight reached different levelsin the populations (random intercepts, Table 1). Al-though there was considerable variation within familiesbetween thermal environments, random slope modelsdid not produce superior model fits, indicating that GxEinteractions were comparatively weak. Infection showedno significant effects on life history traits, probablyowing to the short time period of the infection experi-ment (Table 1).Infection strongly reduced survival (Figure 2). There

was a significant difference between the survival of fishinjected with PBS medium and fish injected with theVibrio sp. isolate (Table 1). Control fish showed lowlevels of mortality towards the end of the experiment.Furthermore, temperature had strong effects on survival(Table 1). Infected fish held at the elevated watertemperature (21°C) died faster (30.7% mortality withinthe first 12 hours of the experiment) and in higher num-bers (54.6% total mortality after 10 days post injection)than infected fish at 17°C (8.8% mortality in the first12 hours, 35.2% total mortality after 10 days post injec-tion). Superior fit of models containing random inter-cepts between populations and random slopes betweenfamilies indicated that also survival rates differed amongpopulations, and that reaction norms of survival ratesvaried significantly among families, thus representingGxE interactions (Table 1).

Effects of thermal stress on quantitative genetic variationWe calculated the drift-based neutral divergence betweenpopulations for all populations combined (global FST =0.00825, range = −0.00055-0.03901) as well as for eachpopulation pair (pairwise FST, Sylt vs. Texel FST = 0.00339,P = 0.19, Sylt vs. Oosterschelde FST = 0.00926, P = 0.01,

3.5

3.6

3.7

3.8

3.9

4.0

stan

dard

leng

th (

cm)

OosterscheldeSyltTexel

0.30

0.35

0.40

0.45

0.50

wei

ght (

g)

OosterscheldeSyltTexel

A B

Figure 1 Effect of thermal stress on life history traits of different stickleback populations. Lines represent means (with standard errors) oflength (A) and weight (B) of three different populations connecting different temperature treatments. Data from all fish were used.


Oosterschelde vs. Texel FST = 0.02187, P = 0.003). We ob-served small, but significant neutral differentiation be-tween the Wadden Sea populations (Sylt and Texel) andthe Oosterschelde. Our FST estimates were somewhatsmaller than previously reported FST values for marinestickleback populations (range of 0.08 to 0.19 [47-49]), butreflected results obtained for other marine fish from theWadden Sea [50].For all investigated traits, quantitative genetic differenti-

ation (QST) exceeded the overall neutral expectation (FST,Figure 3) at 17°C, suggesting directional selection is driv-ing populations to different phenotypic optima at ambienttemperatures [51]. QST values were on average lower at el-evated temperature (21°C, red lines in Figure 3) than at

Table 1 Summary of the effects of thermal stress on life histo

Response Length

DIC

Random Family (F) 1589.39

Population (P) + F 1586.78

P + F (rs) 1604.12

P (rs) + F (rs) 1606.54

Estimate P

Fixed (Intercept) −6.915 0.001

Density −0.084 <0.001

Age 0.034 <0.001

Temperature: 21°C −0.035 0.628

Infection: Non −0.058 0.469

Generalized linear mixed models containing dependent variables length, weight (beffects and density as a covariate. All models contained population and full-siblingrandom factors and were retained when they provided better model fit judged bythe best fitting model. Best fitting models and significant effects are shown in bold

ambient temperature (17°C, blue lines in Figure 3) for alltraits, owing to a trend showing a release of within-population genetic variance of the selected traits (VAnimal,Table 2). This result was most obvious for survival, whereQST values at elevated temperatures did not exceed neu-tral expectations, and divergent selection was only visibleat ambient temperature. Differences in genetic differenti-ation between environments were smaller for weight thanfor length and survival, probably reflecting a stronger en-vironmental influence on this trait that was also visible inthe smaller unit variance components, capturing residualsand environmental variation (Figure 2, Table 2).We found similar slopes of reaction norms between

families and no significant evidence for GxE interactions

ry traits and pathogen resistance

Weight Survival

DIC DIC

1457.55 702.94

1453.26 539.37

1465.05 507.26

1467.54 507.76

Estimate P Estimate P

−8.963 <0.001 0.696 0.573

−0.118 <0.001 0.007 0.895

0.045 <0.001 −0.042 0.467

−0.163 0.004 −1.137 0.011

−0.037 0.583 4.156 <0.001

oth Gaussian) and survival (binomial) with temperature and infection as fixedfamily as random effects. Random slope models (rs) were fitted for boththe Deviance Information Criterion (DIC). Fixed effect tables are based on.

0 50 100 150 200 250

0.0

0.2

0.4

0.6

0.8

1.0

time post injection (hours)

surv

ival

Figure 2 Survival plot of sticklebacks from different temperaturetreatments after infection with Vibrio sp. or PBS medium. Linesshow means for all populations pooled.


for weight and length (Figure 4A, Table 1). This findingwas supported by high and positive genetic correlationcoefficients, that despite being lower than one indicateweak GxE interactions in response to thermal stress(Figure 4B, Table 2). Pathogen resistance, on the otherhand, showed larger variation between environments withmany crossing reaction norms (Figure 4A). Survival alsoshowed strong GxE interactions and significant negativegenetic correlations (Figure 4B, Table 2). Although resistance

0.0 0.2 0.4

QS

surv

ival

leng

th

wei

ght

Figure 3 Quantitative trait and neutral marker differentiation of stickbox represents the distribution range of global FST estimates. Blue lines repat 17°C, while red lines represent 21°C. Dots indicate position of highest delength; only data from infected fish were used for survival.

at ambient temperatures may be only a weak predictorfor resistance under thermal stress, increased resistanceat elevated temperature of some families in all popula-tions (Figure 4A) may suggest that population mean fit-ness may be partly resilient to environmental change.Also, the increase in within-population genetic variationat elevated temperature (Table 2) shows that there isevolutionary potential to facilitate a response to chan-ging climate conditions.Pathogen-induced mortality also exerted correlated se-

lection on life history traits. At ambient temperature, sur-viving fish were significantly longer (parameter estimatelength: 0.649, P <0.001, Figure 5A) and heavier (parameterestimate weight: 0.745, P <0.001, Figure 5B) than fish thatdied. At elevated temperature, differences between surviv-ing and dead fish were smaller on average (parameter esti-mate length: 0.422, P = 0.007, parameter estimate weight:0.382, P = 0.008), and selection differentials for bodylength increased in two out of three cases (Oosterscheldeand Texel), whereas selection differentials for weight de-creased uniformly in all populations.

DiscussionIn this study, we tried to connect the effects of elevatedtemperature on life history traits, pathogen resistance,

0.6 0.8 1.0

T / FST

leback populations at different temperature treatments. The greyresent 95% confidence intervals of QST values for all phenotypic traitsnsities of posterior distributions. All data were used for weight and

Table 2 Estimated variance components of different traits and temperature treatments

17°C 21°C

Weight 17°C VA 0.592 (0.262 – 1.028) Cov 0.862 (0.552 – 1.158)

Vpop 0.216 (0.065 – 1.156) rG 0.791 (0.641 – 0.891)

VU 0.375 (0.128 – 0.578) P rG=0 < 0.001

h2 0.466 (0.137 – 0.713)

21°C VA 0.785 (0.478 – 1.036)

Vpop 0.214 (0.077 – 1.185)

VU 0.167 (0.068 – 0.345)

h2 0.607 (0.298 – 0.794)

Length 17°C VA 0.210 (0.094 – 0.568) Cov 0.386 (0.295 – 0.810)

Vpop 0.269 (0.081 – 1.585) rG 0. 821 (0.534 – 0. 905)

VU 0.698 (0.467 – 0.866) P rG=0 < 0.001

h2 0.183 (0.055 – 0.402)

21°C VA 0.478 (0.264 – 0.946)

Vpop 0.282 (0.097 – 1.699)

VU 0.453 (0.181 – 0.602)

h2 0.345 (0.143 – 0.647)

Survival 17°C VA 0.834 (0.239 – 7.563) Cov −2.655 (−377.98 – -0.387)

Vpop 0.351 (0.055 – 5.407) rG –0.884 (−0.992 – -0.073)

VU 1* P rG=0 < 0.001

h2 ne*

21°C VA 6.604 (0.616 – 100.668)

Vpop 0.701 (0.203 – 4.261)

VU 1*

h2 ne*

Variance components for VAnimal, VPopulation, VUnit (representing residual variation including environmental effects) and resulting heritability (h2) as well as geneticcorrelations across environments (rG) with 95% confidence intervals. Estimates are based on standardized values. *fixed residual variation because categoricaltraits were used and heritabilities were not estimable.


and population-specific potential for evolutionary re-sponses in marine sticklebacks. After only a short acclima-tion period at elevated temperature, we found impacts ofthermal stress on fish weight and especially survival afterinfection. Furthermore, thermal stress tended to releasecryptic within-population genetic variation (VA), whichled to reduced genetic differentiation (QST) between pop-ulations. While life history traits showed positive geneticcorrelations between temperatures and only weak, non-significant GxE interactions, thermal stress led to negativegenetic correlations across environments and strong GxEinteractions in pathogen resistance. These trait-specificpatterns indicate that evolutionary potential of life historytraits (with positive genetic correlations across environ-ments) can be reliably predicted from levels of standinggenetic variation. For survival from infection, on theother hand, negative genetic correlations across envi-ronments indicate that current genetic variation in re-sistance may be a weak predictor for future resistance

evolution, highlighting the crucial importance of diseasein adaptation to new environmental conditions poten-tially favouring pathogens and parasites.In general, thermal stress is likely to affect fish bio-

chemistry, physiology and behaviour [52], and can nega-tively influence fish development and survival [53].Interactions between elevated temperatures and de-creased size and weight have been demonstrated in sev-eral studies [32,54,55], and it is believed that highertemperatures lead to lower food conversion efficiency intemperate sticklebacks, as energy is preferably used tomaintain high metabolic rate [55]. Fish from Sylt wereless affected by thermal stress throughout the experi-ment. This indicates the presence of different pheno-typic optima as well as different thermal tolerancesamong the studied populations. Temperature prefer-ences of sticklebacks for ideal growth were previouslyreported in the range between 15-18°C [32,33], at 19°C[56], and also at 21,7°C [57], showing that thermal

A B

Figure 4 Family-specific reaction norms and genetic correlations across environments. A Family-specific reaction norms of weight, lengthand survival after infection with Vibrio sp. measured in three different stickleback populations. Lines represent different families connecting meansin different temperature treatments. B Character state genetic correlations across thermal environments. All data were used for weight and length;only data from infected fish were used for survival.


optima can differ substantially between populations.Temperate species are able to adapt and shift their ther-mal window through changes in mitochondrial densitiesas well as via other molecular and systemic adjustments[58]. The wide range of thermal optima reported in theliterature in conjunction with our results on quantita-tive genetic differentiation (Figure 3) seem, however, to

suggest that stickleback populations harbour substantialamounts of standing genetic variation and can adaptquickly to local environmental conditions.The reduced sensitivity of the Sylt population can either

be explained by increased tolerance due to phenotypicplasticity, or alternatively, by more favourable experimen-tal conditions matching local conditions when compared

BA

Figure 5 Selection gradients of life history traits under pathogen induced selection. Length (A) and weight (B) of dead fish (red bars = diedat 17°C, blue bars = died at 21°C) and surviving fish (black bars) from different populations. Arrows show the means of the respective distributionsmarking the selection differential S (mean trait survived - mean trait dead) resulting from pathogen-induced mortality. Only data from infectedfish were used for selective gradients.


to the other populations. We can assume that the allo-patric populations from Oosterschelde and Texel werenot only exposed to temperature stress but also had tocope with different water chemistry (our common gar-den experiment used seawater from Sylt). This mighthave led to less favourable overall conditions amplifyingnegative effects of temperature stress, thus resulting in astronger decline of weight at elevated temperature. Al-though we showed that short-term acclimation to increasedtemperature negatively influenced fish development, previ-ous studies have shown that long-term acclimation tohigher temperatures (within the natural range) can poten-tially lead to enhanced development and increased growthrates [58-60]. The Sylt stickleback populations, on the otherhand, continue to show decreased body size when exposedto higher temperatures throughout their lifetime [35], indi-cating that our thermal conditions exceeded their potentialfor acclimation, especially during early phases of develop-ment. Our results presented here thus support our previousresults on a different trait. While we previously found dif-ferences in size during the whole ontogeny [35,61], in thepresent study we did not observe significant influences ofthermal stress on size in later life stages. Rather, we ob-served a response in weight. Although we could not trackindividual fish weight over time (i.e. we did not measureweight before the experimental treatment), we think it ishighly unlikely that we accidentally produced our results bypre-existing bias. Since fish were randomly assigned totreatments, the chances of assigning heavier fish to thecold treatment in 23 out of 31 families used is very small(P = 0.003). Therefore, weight differences more likelyrepresent effects of our treatment than pure chance.Additionally, we have previously shown that the effi-ciency of mitochondrial energy metabolism is reduced

at 21°C [61], leading to higher energy demands and conse-quently lower resource availability. Increased use of re-serves to cope with increased energy demand will lead to adrop in weight, and reduced growth can then be consid-ered a secondary consequence of less available resources.Our experiment also revealed that elevated tempera-

tures have profound impacts on pathogen resistance ofmarine stickleback populations. Infected fish held inwater at 21°C died faster and showed higher overallmortality than infected fish at 17°C. Suboptimal temper-atures may influence innate immunity directly [62] orindirectly as a result of stress-linked overproduction ofimmunosuppressive cortisol [63]. It has also been shownthat elevated temperature negatively affects the resist-ance of fish to diseases by affecting antibody productionand leucocyte activity [12,64]. Either way, stressed or-ganisms are likely to be more susceptible to infectiousdiseases than non-stressed ones [65]. Increased suscep-tibility due to stress can additionally be amplified by in-creased virulence of pathogens at higher temperatures.Vibrios are generally favoured at warmer temperatures[20,21], and the temperature optimum for Vibrio tubia-shii is reported to be at 25°C [66]. We can thus assumethat the elevated temperature of 21°C supported fasterdevelopment and a higher concentration of bacteria inthe fish resulting in higher pathogenicity. Additionally,surviving fish were significantly longer and heavier, withweight reacting faster to environmental change thanlength. Especially for weight, we consistently observeddecreased selection differentials at elevated tempera-tures. The altered response to selection imposed by in-fection might indicate that responses of other fitnesscomponents also can change unpredictably in changingenvironments.


Such changes in responses to selection on single traitswill ultimately also feed back on past and future evolution-ary responses of the host. By using a QST-FST approach, werevealed that, especially for length and survival, the releaseof within-population genetic variance by environmentalstress led to lower QST values (even though between-population variances were not reduced). While the geneticcomponents underlying phenotypic variation were lessdivergent for traits assayed at higher temperature (21°),thermal stress increased genetic variance (VAnimal) andheritability) for most of the selected traits (Table 2). Theeffects of environmental change on quantitative geneticdifferentiation have only rarely been addressed. Hoff-mann and Merilä [67] summarized several outcomes ofhow unfavourable conditions can affect the genetic vari-ation of a trait. First, stressful conditions can increasegenetic variation in traits by increasing rates of recom-bination and mutation [68], by removing low fitness allelesdue to selection [69], or because phenotypic differencesamong genotypes are only expressed as resources becomelimiting [70]. Second, genetic variance and heritabilityare decreased by increased environmental variation [71]or due to limited genetic potential of organisms underpoor nutrition [72]. And third, stressful conditions haveunpredictable effects on the genetic variation of a trait[18,73,74]. Our comparison of indices of population dif-ferentiation in quantitative traits (QST of length, weightand survival) and neutral genetic divergence (FST ofmicrosatellite markers) showed that divergent direc-tional selection was the driving force for reaching differ-ent optima in length, weight and survival at 17°C in ourstickleback populations (QST > FST).Even if the comparison between QST - and FST - values

can reveal the presence and type of selection acting ontraits [75-78], an essential requirement for evolutionarychange is the amount of genetic variability expressed forthe trait under selection [79]. We observed positive geneticcorrelations between character states of life history traitsassayed at different temperatures. Although slopes of thegenetic correlations across environments were significantlydifferent than 1, family-specific reaction norms were ratherflat, indicating that no significant genotype by environmentinteractions (GxE) affected life history traits. In contrast,pathogen resistance showed a negative genetic correlationacross environments. This reversal of genetic effects isprobably due to changing environment-dependent viru-lence mechanisms of the pathogen interacting differentlywith host genotypes, i.e. GPathogenxGHostxE or (GPathogenxE)x(GHostxE) interactions, rather than to pure host genetic ef-fects, i.e. GHostxE interactions. On the other hand, somefamilies within each of the populations showed improvedperformance at elevated temperature, which may lead tohigher fitness in changing environments and adaptation ofthe population as a whole.

The maximal rate of adaptation is of crucial import-ance for estimating the ecological and evolutionary con-sequences of climatic warming. Populations need to shiftthe distribution of phenotypes to maintain optimal fit-ness in the changed environment [14]. Fast adaptationto rising temperature either requires short generationtimes or a substantial amount of standing genetic vari-ation within populations [80]. Even though we foundevolutionary potential in both environments, the nega-tive genetic correlation in pathogen resistance betweenthermal environments demonstrates that adaptation tocurrent conditions can, in certain cases, be a weak pre-dictor for performance in changing environments. Thenegative genetic correlation across environments in sur-vival might also have changed the selection gradients ofassociated life history traits, which might indicate thatdue to pathogen-induced selection, responses of otherfitness components can be variable under unfavourableconditions.

ConclusionOur study shows that thermal stress can negatively im-pact life history traits and pathogen resistance of marinethree-spined sticklebacks. Moreover, environmental stressled to negative genetic correlations across environmentsfor pathogen resistance. Although we found evolutionarypotential in both thermal environments, it is hard to pre-dict the capacity for adaptation by selection of pathogenresistant genotypes from ambient conditions. While posi-tive genetic correlations of life history traits between envi-ronments indicate similar responses to selection, thealtered selective gradient imposed by infection shows thatresponses of vital fitness components can change unpre-dictably in changing environments.

Availability of supporting DataThe data sets supporting the results of this article areavailable in the Pangaea repository http://doi.org/10.1594/PANGAEA.833937 [81].

Competing interestsThe authors declare that they have no competing interests.

Authors' contributionsFSM wrote the manuscript, conducted the fish breeding as well as theexperiment, dissected the fish, and was involved in designing the project,genotyping and performing statistical analyses. LNSS was involved indesigning the project, running pilot studies and writing the manuscript.KMW was involved in designing the project, genotyping, performingstatistical analyses and writing the manuscript. All authors read andapproved the final manuscript.

AcknowledgementsWe would like to thank Nina Eschweiler, Moritz Pockberger, CarolinWendling, René Gerrits, Bert Sinke and David Thieltges for help with fishingin Germany and the Netherlands. René Gerrits and Tobias Mayr for takingcare of the fish. Three anonymous reviewers gave helpful comments onprevious versions of the manuscript. The study was funded by DFG EmmyNoether Programme grant WE 4614/1-1.

http://doi.org/10.1594/PANGAEA.833937



Received: 13 November 2013 Accepted: 14 July 2014Published: 26 July 2014

References1. Hudson PJ, Rizzol IA, Grenfell BT, Heesterbeek H, Dobson AP: The ecology of

wildlife diseases. Oxford: Oxford University Press; 2002.2. Altizer S, Harvell D, Friedle E: Rapid evolutionary dynamics and disease

threats to biodiversity. Trends Ecol Evol 2003, 18(11):589–596.3. Cattadori IM, Haydon DT, Hudson PJ: Parasites and climate synchronize

red grouse populations. Nature 2005, 433(7027):737–741.4. Harvell CD, Mitchell CE, Ward JR, Altizer S, Dobson AP, Ostfeld RS,

Samuel MD: Ecology - Climate warming and disease risks for terrestrialand marine biota. Science 2002, 296(5576):2158–2162.

5. Shiah FK, Ducklow HW: Temperature and substrate regulation of bacterialabundance, production and specific growth-rate in chesapeake bay,USA. Mar Ecol Prog Ser 1994, 103(3):297–308.

6. Karvonen A, Rintamaki P, Jokela J, Valtonen ET: Increasing watertemperature and disease risks in aquatic systems: Climate changeincreases the risk of some, but not all, diseases. Int J Parasitol 2010,40(13):1483–1488.

7. Pounds JA, Bustamante MR, Coloma LA, Consuegra JA, Fogden MPL,Foster PN, La Marca E, Masters KL, Merino-Viteri A, Puschendorf R, Ron SR,Sánchez-Azofeifa GA, Still CJ, Young BE: Widespread amphibianextinctions from epidemic disease driven by global warming.Nature 2006, 439(7073):161–167.

8. Kim K, Harvell CD, Kim PD, Smith GW, Merkel SM: Fungal disease resistanceof Caribbean sea fan corals (Gorgonia spp.). Mar Biol 2000,136(2):259–267.

9. Lafferty KD: The ecology of climate change and infectious diseases.Ecology 2009, 90(4):888–900.

10. Harvell CD, Kim K, Burkholder JM, Colwell RR, Epstein PR, Grimes DJ,Hofmann EE, Lipp EK, Osterhaus AD, nbsp, Overstreet RM, Porter JW, SmithGW, Vasta GR: Emerging Marine Diseases–Climate Links andAnthropogenic Factors. Science 1999, 285(5433):1505–1510.

11. Arnell NW: Climate change and global water resources. Glob EnvironChange-Human Policy Dimens 1999, 9:S31–S49.

12. Alborali L: Climatic variations related to fish diseases and production. VetRes Commun 2006, 30:93–97.

13. Harley CDG, Hughes AR, Hultgren KM, Miner BG, Sorte CJB, Thornber CS,Rodriguez LF, Tomanek L, Williams SL: The impacts of climate change incoastal marine systems. Ecol Lett 2006, 9(2):228–241.

14. Visser ME: Keeping up with a warming world; assessing the rateof adaptation to climate change. Proc R Soc B-Biol Sci 2008, 275(1635):649–659.

15. Eliason EJ, Clark TD, Hague MJ, Hanson LM, Gallagher ZS, Jeffries KM, GaleMK, Patterson DA, Hinch SG, Farrell AP: Differences in Thermal ToleranceAmong Sockeye Salmon Populations. Science 2011, 332(6025):109–112.

16. Perry AL, Low PJ, Ellis JR, Reynolds JD: Climate change and distributionshifts in marine fishes. Science 2005, 308(5730):1912–1915.

17. Gienapp P, Teplitsky C, Alho JS, Mills JA, Merila J: Climate change andevolution: disentangling environmental and genetic responses. Mol Ecol2008, 17(1):167–178.

18. McGuigan K, Nishimura N, Currey M, Hurwit D, Cresko WA: Cryptic geneticvariation and body size evolution in threespine stickleback. Evolution2011, 65(4):1203–1211.

19. Austin B, Austin DA: Bacterial Fish Pathogens: Diseases of Farmed and WildFish. Chichester: Springer; 2007.

20. Thompson JR, Randa MA, Marcelino LA, Tomita-Mitchell A, Lim E, Polz MF:Diversity and dynamics of a north Atlantic coastal Vibrio community.Appl Environ Microbiol 2004, 70(7):4103–4110.

21. Vezzulli L, Previati M, Pruzzo C, Marchese A, Bourne DG, Cerrano C, VibrioSeaC: Vibrio infections triggering mass mortality events in a warmingMediterranean Sea. Environ Microbiol 2010, 12(7):2007–2019.

22. Oberbeckmann S, Fuchs BM, Meiners M, Wichels A, Wiltshire KH, Gerdts G:Seasonal Dynamics and Modeling of a Vibrio Community in CoastalWaters of the North Sea. Microb Ecol 2012, 63(3):543–551.

23. Elston RA, Hasegawa H, Humphrey KL, Polyak IK, Hase CC: Re-emergence ofVibrio tubiashii in bivalve shellfish aquaculture: severity, environmentaldrivers, geographic extent and management. Dis Aquat Org 2008,82(2):119–134.

24. Hada HS, West PA, Lee JV, Stemmler J, Colwell RR: Vibrio tubiashii sp-nov,a pathogen of bivalve mollusks. Int J Syst Bacteriol 1984, 34(1):1–4.

25. Egidius E: Vibriosis-pathogenicity and pathology - a review. Aquaculture1987, 67(1–2):15–28.

26. Austin B, Austin D, Sutherland R, Thompson F, Swings J: Pathogenicity ofvibrios to rainbow trout (Oncorhynchus mykiss, Walbaum) and Artemianauplii. Environ Microbiol 2005, 7(9):1488–1495.

27. Woo PTK, Bruno DW: Fish Diseases and Disorders, Volume 3: Viral, Bacterialand Fungal Infections, 2nd Edition Preface to the. 1st edition; 2011.

28. Barber I, Arnott SA, Braithwaite VA, Andrew J, Huntingford FA: Indirectfitness consequences of mate choice in sticklebacks: offspring ofbrighter males grow slowly but resist parasitic infections. Proc R SocB-Biol Sci 2001, 268(1462):71–76.

29. Kurtz J, Kalbe M, Aeschlimann PB, Haberli MA, Wegner KM, Reusch TBH,Milinski M: Major histocompatibility complex diversity influences parasiteresistance and innate immunity in sticklebacks. Proc R Soc B-Biol Sci 2004,271(1535):197–204.

30. Wegner KM, Kalbe M, Milinski M, Reusch TBH: Mortality selection duringthe 2003 European heat wave in three-spined sticklebacks: effects ofparasites and MHC genotype. BMC Evol Biol 2008, 8:124.

31. Jordan CM, Garside ET: Upper lethal temperatures of threespinestickleback, Gasterosteus aculeatus (L), in relation to thermal andosmotic acclimation, ambient salinity and size. Can J Zool 1972,50(11):1405.

32. Macnab V, Barber I: Some (worms) like it hot: fish parasites grow faster inwarmer water, and alter host thermal preferences. Glob Chang Biol 2012,18(5):1540–1548.

33. Garside ET, Heinze DG, Barbour SE: Thermal preference in relation tosalinity in threespine stickleback, Gasterosteus aculetus L, with aninterpredation of its significance. Can J Zool-Revue Canadienne De Zoologie1977, 55(3):590–594.

34. Sheppard C: Sea surface temperature 1871–2099 in 14 cells around theUnited Kingdom. Mar Pollut Bull 2004, 49(1–2):12–16.

35. Ramler D, Mitteroecker P, Shama L, Wegner K, Ahnelt H: Nonlinear effectsof temperature on body form and developmental canalization in thethreespine stickleback. J Evol Biol 2014, 27:497–507.

36. Schär C, Vidale PL, Lüthi D, Frei C, Häberli C, Liniger MA, Appenzeller C:The role of increasing temperature variability in European summerheatwaves. Nature 2004, 427(6972):332–336.

37. Largiader CR, Fries V, Kobler B, Bakker TCM: Isolation and characterizationof microsatellite loci from the three-spined stickleback (Gasterosteusaculeatus L.). Mol Ecol 1999, 8(2):342–344.

38. Peichel CL, Nereng KS, Ohgi KA, Cole BLE, Colosimo PF, Buerkle CA, SchluterD, Kingsley DM: The genetic architecture of divergence betweenthreespine stickleback species. Nature 2001, 414(6866):901–905.

39. Weir BS, Cockerham CC: Estimating F-statistics for the analysis ofpopulation-structure. Evolution 1984, 38(6):1358–1370.

40. Belkhir KBP, Chikhi L, Raufaste N, Bonhomme F: GENETIX 4.05, logiciel sousWindows TM pour la génétique des populations. In Montpellier. France:Laboratoire Génome, Populations, Interactions, Université de Montpellier II;1996:2004.

41. Whitlock MC, Guillaume F: Testing for spatially divergent selection:comparing QST to FST. Genetics 2009, 183(3):1055–1063.

42. Hadfield JD: MCMC Methods for Multi-Response Generalized LinearMixed Models: The MCMCglmm R Package. J Stat Softw 2010, 33(2):1–22.

43. R Development Core Team: R: A language and environment for statisticalcomputing R Foundation for Statistical Computing; 2008.

44. Spiegelhalter DJ, Best NG, Carlin BR, van der Linde A: Bayesian measuresof model complexity and fit. J R Stat Soc Series B-Stat Methodol 2002,64:583–616.

45. Spitze K: Population-structure in Daphnia-Obtusa - quantitative geneticand allozymic variation. Genetics 1993, 135(2):367–374.

46. Falconer DS: The Problem of Environment and Selection. Am Nat 1952,86(830):293–298.

47. Leinonen T, Cano JM, Makinen H, Merila J: Contrasting patterns of bodyshape and neutral genetic divergence in marine and lake populations ofthreespine sticklebacks. J Evol Biol 2006, 19(6):1803–1812.

48. Raeymaekers JAM, Van Houdt JKJ, Larmuseau MHD, Geldof S, VolckaertFAM: Divergent selection as revealed by P-ST and QTL-based F-ST inthree-spined stickleback (Gasterosteus aculeatus) populations along acoastal-inland gradient. Mol Ecol 2007, 16(4):891–905.

49. Reusch TBH, Wegner KM, Kalbe M: Rapid genetic divergence in postglacialpopulations of threespine stickleback (Gasterosteus aculeatus): the role


of habitat type, drainage and geographical proximity. Mol Ecol 2001,10(10):2435–2445.

50. Hoarau G, Rijnsdorp AD, Van der Veer HW, Stam WT, Olsen JL: Populationstructure of plaice (Pleuronectes platessa L.) in northern Europe:microsatellites revealed large-scale spatial and temporal homogeneity.Mol Ecol 2002, 11(7):1165–1176.

51. Merila J, Crnokrak P: Comparison of genetic differentiation at marker lociand quantitative traits. J Evol Biol 2001, 14(6):892–903.

52. Beitinger TL, Bennett WA, McCauley RW: Temperature tolerances of NorthAmerican freshwater fishes exposed to dynamic changes in temperature.Environ Biol Fish 2000, 58(3):237–275.

53. Dominguez M, Takemura A, Tsuchiya M, Nakamura S: Impact of differentenvironmental factors on the circulating immunoglobulin levels in theNile tilapia, Oreochromis niloticus. Aquaculture 2004, 241(1–4):491–500.

54. Donelson JM, Munday PL, McCormick MI, Pitcher CR: Rapid transgenerationalacclimation of a tropical reef fish to climate change. Nat Clim Chang 2012,2(1):30–32.

55. Guderley H, Leroy PH: Family origin and the response of threespinestickleback, Gasterosteus aculeatus, to thermal acclimation. J CompPhysiol B-Biochem Syst Environ Physiol 2001, 171(2):91–101.

56. Allen JRM, Wootton RJ: The effect of ration and temperature on thegrowth of the 3-spined stickleback, Gasterosteus aculeatus L. J Fish Biol1982, 20(4):409–422.

57. Lefebure R, Larsson S, Bystrom P: A temperature-dependent growth modelfor the three-spined stickleback Gasterosteus aculeatus. J Fish Biol 2011,79(7):1815–1827.

58. Portner HO: Climate change and temperature-dependent biogeography:oxygen limitation of thermal tolerance in animals. Naturwissenschaften2001, 88(4):137–146.

59. Green BS, Fisher R: Temperature influences swimming speed, growth andlarval duration in coral reef fish larvae. J Exp Mar Biol Ecol 2004,299(1):115–132.

60. Sponaugle S, Grorud-Colvert K, Pinkard D: Temperature-mediated variationin early life history traits and recruitment success of the coral reef fishThalassoma bifasciatum in the Florida Keys. Mar Ecol Prog Ser 2006,308:1–15.

61. Shama L, Strobel A, Mark F, Wegner K: Transgenerational plasticity inmarine sticklebacks: maternal effects mediate impacts of a warmingocean. Funct Ecol 2014, doi:10.1111/1365-2435.12280.

62. Nikoskelainen S, Bylund G, Lilius EM: Effect of environmental temperatureon rainbow trout (Oncorhynchus mykiss) innate immunity. Dev CompImmunol 2004, 28(6):581–592.

63. Weyts FAA, Cohen N, Flik G, Verburg-van Kemenade BML: Interactionsbetween the immune system and the hypothalamo-pituitary-interrenalaxis in fish. Fish Shellfish Immunol 1999, 9(1):1–20.

64. Landis SH, Kalbe M, Reusch TBH, Roth O: Consistent Pattern of LocalAdaptation during an Experimental Heat Wave in a Pipefish-TrematodeHost-Parasite System. PLoS One 2012, 7:1.

65. Lafferty KD, Holt RD: How should environmental stress affect thepopulation dynamics of disease? Ecol Lett 2003, 6(7):654–664.

66. Temperton B, Thomas S, Tait K, Parry H, Emery M, Allen M, Quinn J,MacGrath J, Gilbert J: Permanent draft genome sequence of Vibriotubiashii strain NCIMB 1337 (ATCC19106). Stand Genomic Sci 2011,4(2):183–190.

67. Hoffmann AA, Merila J: Heritable variation and evolution under favourableand unfavourable conditions. Trends Ecol Evol 1999, 14(3):96–101.

68. Hoffmann AA, Parsons PA: Evolutionary genetics and environmental stress. 1991.69. Kawecki TJ, Barton NH, Fry JD: Mutational collapse of fitness in marginal

habitats and the evolution of ecological specialisation. J Evol Biol 1997,10(3):407–429.

70. Hartl DL, Dykhuizen DE, Dean AM: Limits of adaptation - the evolution ofselective neutrality. Genetics 1985, 111(3):655–674.

71. Blum A: Plant Breeding for Stress Environments: CRC Press. 1988.72. Gebhardthenrich SG, Vannoordwijk AJ: Nestling growth in the great tit 1.

Heritability estimates under different environmental conditions. J EvolBiol 1991, 4(3):341–362.

73. Ebert D, Yampolsky L, Stearns SC: Genetics of life-history in daphnia-magna.1. Heritabilities at two food levels. Heredity 1993, 70:335–343.

74. Gavrilets S, Scheiner SM: The genetics of phenotypic plasticity 6.theoretical predicitions for directional selection. J Evol Biol 1993,6(1):49–68.

75. Knopp T, Cano JM, Crochet P-A, Merila J: Contrasting levels of variation inneutral and quantitative genetic loci on island populations of moor frogs(Rana arvalis). Conserv Genet 2007, 8(1):45–56.

76. Leinonen T, O'Hara RB, Cano JM, Merila J: Comparative studies ofquantitative trait and neutral marker divergence: a meta-analysis. J EvolBiol 2008, 21(1):1–17.

77. Roberge C, Guderley H, Bernatchez L: Genomewide identification of genesunder directional selection: Gene transcription Q(ST) scan in divergingAtlantic salmon subpopulations. Genetics 2007, 177(2):1011–1022.

78. Tubiash HS, Chanley PE, Leif-Son E: Bacillary necrosis, a disease of larvaland juvenile bivalve mol-lusks. I. Etiology and epizootiology. J Bacteriol1965, 90(4):1036–1044.

79. Charmantier A, Garant D: Environmental quality and evolutionarypotential: lessons from wild populations. Proc R Soc B-Biol Sci 2005,272(1571):1415–1425.

80. Poertner HO, Farrell AP: Ecology Physiology and climate change. Science2008, 322(5902):690–692.

81. Schade FM, Shama LNS, Wegner KM: Data from: Impact of thermal stresson evolutionary trajectories of pathogen resistance in three-spinedstickleback (Gasterosteus aculeatus). Pangaea 2014, http://doi.org/10.1594/PANGAEA.833937.

doi:10.1186/s12862-014-0164-5Cite this article as: Schade et al.: Impact of thermal stress onevolutionary trajectories of pathogen resistance in three-spinedstickleback (Gasterosteus aculeatus). BMC Evolutionary Biology 2014 14:164.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit



impact of thermal stress on evolutionary trajectories of ... · daily ad libitum with live artemia...

Documents